Data driven model free adaptive iterative learning perimeter control for large-scale urban road networks

Most perimeter control methods in literature are the model-based schemes designing the controller based on the available accurate macroscopic fundamental diagram (MFD) function with well known techniques of modern control methods. However, accurate modeling of the traffic flow system is hard and time-consuming. On the other hand, macroscopic traffic flow patterns show heavily similarity between days, and data from past days might enable improving the performance of the perimeter controller. Motivated by this observation, a model free adaptive iterative learning perimeter control (MFAILPC) scheme is proposed in this paper. The three features of this method are: (1) No dynamical model is required in the controller design by virtue of dynamic linearization data modeling technique, i.e., it is a data-driven method, (2) the perimeter controller performance will improve iteratively with the help of the repetitive operation pattern of the traffic system, (3) the learning gain is tuned adaptively along the iterative axis. The effectiveness of the proposed scheme is tested comparing with various control methods for a multi-region traffic network considering modeling errors, measurement noise, demand variations, and time-changing MFDs. Simulation results show that the proposed MFAILPC presents a great potential and is more resilient against errors than the standard perimeter control methods such as model predictive control, proportional-integral control, etc.